прагматика и медиа дискурс / Structures of discourse and structures of power

actions, and may have recourse to a rhetoric of dramatic or emotional appeals, or to various forms of topical or stylistic originality. The power groups involved here form what we called the symbolic elites. A specific case of this class of discourses is news reports in the media, which not only describe current events and their possible consequences, but which essentially portray the actions, and represent the opinions of the political, economic, military, and social power elites. It is mainly in this way that the consensual basis of power is manufactured, and how the general public gets to know who has power and what the powerful want. This is a crucial condition for the development of the supporting ideological framework of power, but also for various forms of resistance ("know thine enemies").

This first typology shows that the discursive enactment of power is mostly persuasive. Powerful groups or institutions only rarely have to prescribe what the less powerful should do, although ultimately such directives may be decisive in controlling others, as is especially the case in state control. Rather, they argue by providing economic, political, social, or moral reasons, and by managing the control of relevant information. In this way, communication may be biased through selective release of information that is favorable to the power elites, or by constraining information that is unfavorable to them. The realization of these goals may be facilitated by various rhetorical or artistic means.

Levels of Discourse and Power

A second dimension goes beyond this simple typology of discourse genres and their contributions to social control. It features the various levels of discourse that may specifically enact, manifest, express, describe, signal, conceal, or legitimate power relations between discourse participants or the groups they belong to.

Thus as we have seen earlier, power may first be enacted at the pragmatic level through limited access, or by the control of speech acts, such as commands, formal accusations, indictments, acquittals, or other institutional speech acts. Second, in conversational interaction, one partner may control or dominate turn allocation, self-presentation strategies, and the control of any other level of spontaneous talk or formal dialogue. Third, selection of discourse type or gene may be controlled by more powerful speakers, for instance in the classroom, courtroom, or within the corporation: Sometimes stories of personal experiences are allowed, but more often than not, they tend to be censored in favor of the controlled discourse genres of the business at hand, for instance interrogations. Fourth, outside of everyday conversation, topics are mostly controlled by the Tules of the communicative situation, but their initiation, change, or variation are usually controlled or evaluated by the more powerful speaker. The same is true for style and rhetoric.

Dimensions of Power

The analysís of power structures allows us to list other relevant categories, namely, those dimensions of power that may have an ímpact on díscourse and its structures: The various institutions of power, the internal power structures of these institutions, power relations between different social groups, and the scope or domain of the exercise of power by (members of) these institutions or groups. Without a further analysis of these structures and dimensions of social power, we simply argue here that they are also manifested in the various structures of "powerful" text and talk.

In this list we first find the major power institutions, such as the government, parliament, state agencies, the judiciary, the military, big corporations, the political parties, the media, the unions, the churches, and the institutions of education. Each of these institutions may be associated with its specific discourse genres, communicative events, topics, styles, and rhetorics. Second, there is the usual hierarchy of position, rank, or status within these institutions and these imply different speech acts, genres, or styles, for example, those signaling authority and command.

Third, parallel and sometimes combined with the institutions, we have, group power relations, such as those between the rich and the poor, men and women, adults and children, white and black, nationals and foreigners, the highly educated and those who have little education, heterosexuals and homosexuals, believers and nonbelievers, the moderates and the radicals, the healthy and the sick, the famous and the unknown, and generally those between Us and Them. Both within institutional and in everyday, informal interaction, these power relations may be structurally enacted by the members of the respective dominant groups. As is the case for institutional members, members of dominant groups may derive their individually exercised power from the overall power of the group they belong to. The effect on discourse in these cases will be especially obvious in the unbalanced control of dialogue, turn taking, speech acts, topic choice, and style.

Fourth, the enactment of power may be analyzed as to its domain of action or scope and type of influence. Some institutions or their leading members may accomplish discursive acts that affect whole nations, states, cities, or large organizations, or they may affect life and death, health, personal freedom, employment, education, or the private lives of other people, whereas other institutions or their members have a less broad and a less serious impact on other people.

Finally, we may distinguish between the various kinds of legitimacy for these forms of social control, which rnay vary between total control imposed or maintained by force (as in a dictatorship, and in some domains also in a democratic system of government), on the one hand, and partial control sanctioned by an elite, by a majority, or on the other hand, by a more or less

general consensus. These (gradual) differences reflect the possible sanctions of the powerful, as well as the acceptance or resistance of those subjected to the enactment of power.

These differences in the modes of legitimation are also manifest in different genres, topics, and styles of discourse. Discussion, argumentation, and debate, for example, are not characteristic of dictatorial discourse. Hence the importance of the amount and nature of discursive legitimation in these different sorts of power systems. It may be expected that each political system, viewed as an institutionalization of power, for instance by the state, is associated with its own characteristic orders or modes of discourse. Since the principles (norms, rules, values, goals) of legitimacy are embedded in an ideology, the processes of legitimation will also appear as discursive processes.

Different Approaches

With these various dimensions of power in mind, we should be able to make the next step and establish systematic links between these dimensions and the various structural dimensions of discourse. However, this may be done in different ways and from different, complementary perspectives. Thus the social scientist may start with an analysis of the dimensions of social power just mentioned and then examine through what discourses or discursive properties these power structures are expressed, enacted, or legitimated. This (macro) approach favors a more general and integrated analysis of various discourse genres and properties related to a class, institution, or group (for instance, the discourse of the legal system, or the patriarchal power of men over women). On the other hand, the sociolinguist will usually start with an analysis of specific properties of language use or discourse, and try to show how these may vary, or depend on, different social positions, relations, or dimensions, for example, those of class, gender, ethnic group, or situation. This perspective will usually pay more detailed attention to linguistic properties of text and talk, and take a more general view of the various social "circumstances" of such properties.

We opt for an approach that combines the advantages of these two alternatives, namely the analysis of discursive (sub)genres and communicative events in social situations (Brown & Fraser, 1979). Such a "situation analysis" requires an integration of both discourse analysis and social analysis. Through an interdisciplinary study of everyday conversations, classroom dialogues, job interviews, service encounters, doctors' consultations, court trials, boardroom meetings, parliamentary debates, news reporting, advertising, or lawmaking, among many other communicative events, we are able to assess both the relevant discourse structures and the relevant structures of dominance and control in the social context. That is, understanding these communicative genres requires an analysis of participant representation, interactional strategies, turn allocation, topic and code selection, stylistic registers, rhetorical operations, and also an analysis of the roles, relations,

rules, norms, or other social constraints that govern the interaction of participants as social group members. In this way, we capture both the properties and processes of text and talk, and the micromechanisms of social interaction and societal structure. Also, this level and scope of analysis allows a sociocognitive assessment of knowledge, opinions, attitudes, ideologies, and other social representations that exercise the cognitive control of acting agents in such situations. Finally, these social microstructures (e.g., the lesson) may in turn be related (e.g., by comparison or generalization) to relevant social macrostructures, such as institutions (e.g., the school, the education system, and their ideologies) and overall social relations (e.g., the dominance of whites over blacks) (Knorr-Cetina & Cicourel, 1981).

POWER IN DISCOURSE: A REVIEW

In the previous sections, I have given a brief theoretical analysis of the notion of power and its links with discourse and communication. We have witnessed how the powerful have recourse to many strategies that allow them to control the material and symbolic production of text and talk, and, therefore, part of the cognitive processes that underlie the cognitive management and the manufacturing of consent from the less powerful. On several occasions, this discussion has mentioned some properties of discourse that are specifically affected by this process of (re)productive control, for instance, conversational turn taking, topics, and style. In the remainder of this chapter, we analyze in more detail how power is actually expressed, signaled, reproduced, or legitimated in various structures of text and talk. Whereas the previous sections focused on various social strategies of discourse and communication control, we will now systematically examine the discursive strategies that implement such (inter)actions, and briefly review empirical studies that show power "at work" in text and talk. We will organize our discussion around a few selected discourse types, namely, subgenres or communicative events, that also embody typical social relations, including specific power relations. In this discussion, a reinterpretation of research will sometimes be necessary, for instance, when the notion of power is not used as such. We begin with various sorts of spoken, dialogical discourse, and then discuss written types of text. We will focus on social power and disregard types of individual power, influence, or status in interpersonal communication (see, Berger, 1985 for a review of this work, and Brooke & Ng, 1986, and Falbo & Peplau, 1980, for empirical studies on interpersonal influence).

Conversation

Although the analysis of conversation generally presupposes that speakers have equal social roles (Sacks, Schegloff, & Jefferson, 1974; Atkinson & Heritage, 1984; McLaughlin, 1984), it is obvious that group and institutional

membership of speakers, and in general social inequality, introduce differences in control over the ongoing dialogue. These differences appear, for instance, in talk between men and women, adults and children, whites and blacks, the rich and the poor, or between the more or less educated. It is assumed that such control by the more powerful speaker may extend to turn allocation or appropriation, speech act choice, topic selection and change, and style. The enactment of this control, however, need not be static, but may be dynamically negotiated or challenged by the less powerful speakers. In other words, talk is continuously contextualized by signaling various conditions or constraints of the social situation in general, and by the social relationships between the speech participants, in particular. And although it makes sense to make a distinction between everyday, personal, or informal talk, on the one hand, and, on the other hand, formal, institutional discourse, it should be stressed that informal or private discourse may be imbued with formal and institutional constraints. Conversely, institutional discourse also may be informal and an everyday accomplishment among other social practices.

Conversation Between Parents and Children

One of the more obvious power differences in many cultures is that between parents and children. Although there is important cultural variation (Snow & Furgeson, 1977), and differences between fathers and mothers (Gleason & Geif, 1986), parental control is generally enacted in parent-child talk in many ways: "The low status of children in stratified societies can keep them silent, forbid them to initiate or discuss certain topics, prevent them from interrupting, or require them to use a special deferential variety of speech" (Ervin-Tripp & Strage, 1985, p. 68).

As these and other authors show in detail, parents may also control child behavior more directly, for example, through scolding, threatening, directing, or correcting children in talk. More indirect forms of action control in parent-child talk may take the form of advice, requests, or inducement through promises. These differences in parental control in talk have often been related to class differences (Cook-Gumperz, 1973). Relevant to our discussion of social power, social representations of power are acquired and displayed rather early, as through different forms of discursive politeness and deference, or through verbal power play and ritual (Bavelas, Rogers, & Millar, 1985; Ervin-Tripp, O'Connor, & Rosenberg, 1984; Labov, 1972; Lein & Brenneis, 1978).

Conversation Between Women and Men

The power differences between women and men and their manifestation in language have received extensive attention, especially during the last decade, and by feminist researchers (Eakins & Eakins, 1978; Kramarae, 1980, 1983; Spender, 1980; Thorne & Henley, 1975; and Thorne, Kramarae, & Henley,

1983, who provide an extensive bibliography). Therefore, we mention only a few general conclusions of this important work, which in many respects has beco me paradigmatic for the analysis of power in language and communication, and focus on the more recent studies of gender power in discourse (for a brief review, see West & Zimmerman, 1985).

Although differences may sometimes be subtle and dependent on situation (Leet-Pellegrini, 1980), and on social position (Werner, 1983), it has been found that women generally "do more work" than men do in conversation, by giving more topical support, by showing more interest, or by withdrawing in situations of conflict (Falbo & Peplau, 1980; Fishman, 1983). Several studies document that men tend to interrupt women more often, especially at irregular turn transition places (Eakins & Eakins, 1978; Natale, Entin, & Jaffe, 1979; West & Zimmerman, 1983).

Some of the studies collected by Trómel-Plótz (1984) show that male dominance is not restricted to informal situations, such as the home, but also appears in public contexts, such as television talk shows, which are moderated mostly by men (see also Owsley & Scotton, 1984). For instance, women tend to get the floor less often than men do, and men talk longer, more often, and use long, complicated sentences and various types of pseudostructuring of conversational contributions.

Gender differences in talk may also be studied in a more general perspective as instances of "powerful" and "powerless" speech, which may be found in other social situations (Bradac & Street, 1986; Erickson, Lind, Johnson, & O'Barr, 1978), to which we tum next.

Racist Talk

What is true for the subordination of women in talk, also holds for discourse addressed to, or about blacks and other minority groups in many Western countries (Smitherman-Donaldson & van Dijk, 1987). White group power may also be exercised through verbal abuse and derogation of minority group members (Allport, 1954). Although there are many historical and literary sources that document the pervasiveness of racial slurs, there are few systematic studies of their usage and functions. Kennedy (1959) provides a brief list of "etiquette rules" for the ways blacks and whites should address each other in the period of Jim Crow racism in the United States. One of these rules was that blacks should never be addressed as "Mr.," "Mrs.," "Sir," or "Ma'am," but by first names only, whereas whites always must be addressed in the polite form. Although the last decades have seen much of this verbally expressed racism mitigated because of changing official norms and laws, racial slurs still exist in everyday white talk. Verbal derogation of blacks, as well as of Chinese, Italian, Mexican, or Puerto Rican Americans is common in the United States, and of Turkish, Moroccan, South Asians, Caribbean, and other minorities or immigrants in Western Europe (Helmreich, 1984).

Ethnic conflict may also be manifested in different speech styles that lead to misunderstanding and stereotyping (Kochman, 1981). Within a German project on language acquisition by immigrant workers, attention was paid to the ways these "Gastarbeiter" were addressed in terms of a perceived, simplified "foreigner German" (Dittmar & Stutterheim, 1985; Klein & Dittmar, 1979). Often, such talk by itself may signal superiority of the speakers and their group. This is an interesting specific case of the functions of linguistic accommodation and conflict in interethnic communication (Giles & Powesland, 1975; Giles & Smith, 1979; Gumperz, 1982a, 1982b).

Much recent research on prejudice and racism suggests that even if racist opinions, talk, and action have become more indirect and subtle in certain contexts, basic attitudes may not have changed very much (Barker, 1981; Dovidio & Gaertner, 1986; Essed, 1984). Greenberg, Kirkland, and Pyszczynski (1987) show that the use of racial slurs by experimental confederates against black subjects may activate such basic attitudes among white subjects and result in more negative evaluations of these black subjects. Among the conservative elites, racist discourse has taken a more "cultural" orientation during the last decade. Such discourse emphasize assumed cultural differences between in-groups and out-groups, and sometimes subtly advocates nationalist cultural autonomy of the dominant white group (Seidel, 1987a, 1987b).

In my own work on the expression of ethnic opinions and prejudice in everyday talk, such explicit racial slurs appear to be rare, both in the Netherlands and in California (van Dijk, 1984a, 1987a). However, the informal interviews on which my research is based are typically examples of talk with relative strangers (university students), and, therefote, such talk is likely to be heavily monitored by official norms of nondiscrimination. In fact, white people routinely express their knowledge of such norms, and elaborately affirm that whatever they may say about "foreigners" they do not mean to be racists.

Therefore, the overall strategy of talk about minorities is twofold. On the one hand, many white people express negative experiences and opinions about ethnic minority groups. On the other hand, however, this negative "other-presentation" is systematically balanced by positive self-presentation, namely, as tolerant, nonracist, understanding citizens. This overall strategy is implemented by many local strategies and tactics, such as apparent denials and concessions ("I have nothing against them, but ." "There are also good ones among them, but . . ." and so on), contrasts that emphasize group differences, competition, generally the us/ them opposition ("We work hard, and they don't have to do anything"), and transfer ("I don't mind, but other people in the country, city, street, or department do"). Besides such semantic and rhetorical strategies of positive self-presentation, negative other-presen- tation is mainly implemented by argumentation and concrete storytelling. Stories are based one's own personal experiences, and, therefore, "true" and good "evidence" for negative conclusions. Most of these stories feature events

and actions of minority groups that are perceived to violate dominant (white) norms, values, goals, and interests, but which also happen to substantiate current stereotypes and prejudices. Often, the news media are used to legitimate such stories and opinions, for instance by referring to minority crime "about which you read in the paper everyday." More subtly, conversational properties such as hesitations, repairs, and corrections provide insight into the underlying cognitive processes and monitoring in such talk. Lexical choice and the use of identifying pronouns and demonstratives also suggest social distance: "them," "those people," "those Turks (Mexicans, and so on)." In this way, everyday talk among white majority group members reproduces such prejudices within the ingroup, while at the same time verbally confirming group membership, and group goals and norms, which in turn are relevant in the maintenance of white group power.

Institutional Dialogue

Dialogues with and within institutions or organizations are forms of institutional interaction, and, therefore, also enact, display, signal, or legitimate a multitude of power relations (Pettigrew, 1973; Pfeffer, 1981). Participants in such interactions may follow context dependent rules and norms of interaction, but may also negotiate different roles or positions, including those of status, hierarchy, or expertise. Another difference with everyday, informal conversation is that institutional members are mostly professionals, experts "at work" (see also Coleman, 1984, 1985b). Let us examine some of the more prominent subgenres of institutional dialogue.

Job Interviews

Ragan (1983) showed that in job interviews power differences manifest themselves in what she calls "aligning actions," such as accounts, metatalk, side sequences, digressions, or qualifiers. Interviewers more often had recourse to strategies that control conversational pace and progress, such as formulations, metatalk, and metacommunicative digressions. Applicants, on the contrary, are more often engaged in justifying or explaining their behavior, for instance through accounts, qualifiers, and "you knows," even when these were unnecessary. This study complements earlier social psychological work on the (power) effect of language attitudes in job interviews, which shows that otherwise identical applicants may be discriminated against because of their foreign accent, for instance, by getting lower evaluations for higher-level jobs and higher evaluations for lower-level jobs (Kalin & Rayko, 1980).

In a series of experimental studies, Bradac and associates examined the role of powerful and powerless styles in job interviews (Bradac & Mulac, 1984). As in early studies of women's language, hesitations, and tag questions were found to characterize the powerless style (see also Bradac & Street, 1986). We

shall see that similar results have been found in styles of courtroom talk.

Doctor-Patient Discourse

Doctor-patient discourse is just one specific example of medical discourse in general (Fisher & Todd, 1983, 1986; Freeman & Heller, 1987), which has often been criticized for a variety of reasons, including the abuse of power by medical practitioners. Edelman (1974), in a critical article, shows how the language of people in the helping professions, typically in psychiatry, in many ways conceals the real nature of their intentions and actions, which are geared toward the control of patients. In this way, direct power may be masked by the discourse of "helping," in which patients who have good reasons to be angry may be categorized as "aggressive." Such patients will be put in what is euphemistically called a "quiet room" instead of "solitary confinement." Similarly, the use of such terms as "predelinquent" may mean that professionals get carte blanche in the "treatment" of (mostly powerless, e.g., young, poor) people who have shown no sign of deviance. Professional power here combines with the power of class and age. Indeed, as we shall see next, power seldom comes alone: Institutional power is frequently enacted at the same time as group power derived from gender, class, race, age, subculture, or nationality (see also Sabsay & Platt, 1985).

West (1984) shows that the inherent social asymmetry in doctor-patient relationships is also displayed in their conversations, and that gender and race play a role here: Male doctors interrupt patients (especially black patients) much more often than the reverse, without any medical function or relevance; on the contrary, these interruptions make them miss important information. Female doctors, however, are interrupted more often by their (male) patients. Generally, in doctor-patient talk there is an imbalance in information exchange: Doctors initiate most questions and patients stutter when asking their few questions, with the exception of a specific type of conditional query. West concludes that, "Quantitative and qualitative evidence suggests that physicians stand in nearly godlike relation to their patients—as entities `not to be questioned'" (West, 1984, p. 51). Formal expressions are used to address the doctor, whereas doctors tend to use the first names of patients, especially when the patients are black. Fisher and Todd (1983) also find an interaction between medical and gender power. They showed that female patients are subject to "friendly persuasion" by (male) practitioners to use birth control pills, while being kept uninformed about the pills' possible negative effects or about alternative forms of birth control.

In a critical analysis of clinical interviews, Mishler (1984) found discursive evidence for the domination of what he calls the "biomedical voice" of doctors, and concludes: "Typically, the voice of the lifeworld was suppressed and patients' efforts to provide accounts of their problems within the context of their lifeworld situations were disrupted and fragmented" (p. 190). Treichler, Frankel, Kramarae, Zoppi, and Beckman (1984) argue that the

physician's focus on biomedical aspects hinders the full expression of the patient's concerns. Thus concerns readily expressed to a medical student were not included in the physician's medical records. Doctors are found to use irony in showing dismissal of the patient's complaints. Finally, just as for job interviews, social psychological work on language attitudes shows that doctors may evaluate their patients differently depending upon whether or not they have a dialect or sociolect accent (Fielding & Evered, 1980).

What has been found for general practitioners may be expected to be true for other medical professionals. Coleman and Burton (1985) studied control in dentist-patient consultations in Great Britain, and found that dentists control both verbal and nonverbal activity: Dentists talk 71% and patients 26% of the time, (assistants 3%). Dentists have more turns, and longer turns (4.6 versus 2.1 seconds). Obviously, control in this case takes a very literal form: Patients usually have their mouths open, but are still prevented from speaking in such a situation, and, therefore, have little to say in the first place. Compliance with dentists' power may also depend on fear of pain. Thus the authors found that dentists regularly respond to patients' reports by making no acknowledgment, by minimizing them as irrelevant, or by dismissing them as incorrect. As is the case for most professional forms of power, the majos resource of dentist dominance is expertise (see also Candlin, Burton, & Coleman, 1980).

As noted earlier, power may derive from institutional organization and routinization. Medical power is a characteristic example. The results of the studies just reviewed should also be interpreted in that perspective. Thus Strong (1979) specifies some other factors that limit the freedom of patients in consultation discourse: Doctors use technical language (see also Coleman, 1985a); there are few doctors and many patients; doctors are organized and patients are usually not; doctors have high status; in some countries, there are no or few (affordable) alternatives for the public health service provided by doctors, and, therefore, little medical competition and reduced possibilities for second opinions. We see that the local enactment and organization of power in doctor-patient talk is intricately interwoven with more general social and institutional forms of control.

These findings are also relevant in counsefing or admission interviews, in which professionals act as gatekeepers of institutions and may exert relevant group power on the differential conversational treatment of minority clients or candidates (Erickson & Shultz, 1982; Mehan, 1986). Similarly, in classroom talk, teachers may be expected to exercise control over students through a series of strategies: They decide about discourse type, they initiate and evaluate topics and question-answer sequences, they monitor student speech style, and generally control both the written and spoken discourses of the students. Unfortunately, although there is much work on classroom dialogues (Sinclair & Brazil, 1982; Stoll, 1983; Wilkinson, 1982), little specific attention is paid to there routine enactments of institutional power.